Standard deviation is a measure of how widely observations are dispersed around their mean. It is the square root of variance, so it is expressed in the same units as the data rather than in squared units. A larger standard deviation means the observations are more spread out; a smaller value means they cluster more tightly around the mean.
The denominator changes the estimate
For a complete population, standard deviation uses the average squared deviation from the population mean and then takes the square root. When a sample is used to estimate population variability, a common sample formula divides the sum of squared deviations by n − 1 before taking the square root: s = √[Σ(xᵢ − x̄)²/(n − 1)]. The denominator is not a cosmetic choice. Using n instead of n − 1 produces a different value, with the difference most noticeable in short samples. For an independent random sample from a population with finite variance, the n − 1 correction makes the sample variance unbiased; taking the square root does not generally produce an unbiased estimator of the population standard deviation.
When dispersion is used as volatility
In market research, standard deviation is often applied to returns to quantify realized dispersion over a chosen window. That number is inseparable from its construction: daily versus weekly observations, arithmetic versus logarithmic returns, lookback length, missing-data treatment and the denominator convention can all change the result. Standard deviation calculated on price levels answers a different question; it measures dispersion of those price observations around their mean and should not automatically be interpreted as return volatility.
Because deviations are squared, extreme observations receive disproportionate weight. That makes standard deviation responsive to large market moves but also sensitive to outliers and heavy tails. It does not imply that the underlying distribution is normal, nor does one standard deviation carry a fixed probability meaning without a distributional assumption. A volatility regime shift can also make a single full-sample value conceal important time variation.
For reproducible market research, the important point is therefore not merely the reported standard deviation but what was measured, over which window, at what frequency and with which denominator. Those choices should be aligned before values are compared across markets, models or software implementations.